evolutionary algorithm can be combined with on-policy?

[Marioooooooooooooo]

Congratulations to you for your valuable work on evolutionary reinforcement learning. I want to know whether evolutionary algorithm can be combined with on-policy(for example, PPO) algorithm, if there are relevant references, please point out, thank you very much.

[yeshenpy]

Sure! Some works combine Evolutionary Algorithms with on-policy RL, such as Supe-RL, G2N, etc. More work will be incorporated here later and updated on this website. Hopefully, it can address your problem.

[1]. Genetic soft updates for policy evolution in deep reinforcement learning.
[2]. Genetic-gated networks for deep reinforcement learning.